Performance Of Speaker Verification System Research Articles

Sinc‐attention feature extraction for trivial‐event based speaker verification

AbstractHuman speech contains trivial events that are non‐subjectively controlled and are limited by physical characteristics. These events are represented by a series of extremely short utterances, such as the ‘hmm’ phrase used for expressing doubt or confirmation. In speaker verification (SV), such events can be a robust choice to verify the real speaker from disguised speech since they are less affected by the randomness of pronunciation. However, trivial events like ‘hmm’ contain little linguistic information and are extremely short, thus the performance of SV systems will decrease drastically. In this letter, a Sinc‐Attention feature extraction method is proposed to extract more discriminative speech features from speech signals to achieve a robust SV system for trivial events. Learnable filters are utilized to obtain low‐level representations of the speech for SV. Moreover, a novel adaptive weighting of features inspired by attention mechanisms is proposed, which effectively improves the representation capability of speaker features. The experiments on different models prove that our method helps SV systems achieve a lower equal error rate (EER) than hand‐crafted feature‐based systems.

DeltaVLAD: An efficient optimization algorithm to discriminate speaker embedding for text-independent speaker verification

<abstract> <p>Text-independent speaker verification aims to determine whether two given utterances in open-set task originate from the same speaker or not. In this paper, some ways are explored to enhance the discrimination of embeddings in speaker verification. Firstly, difference is used in the coding layer to process speaker features to form the DeltaVLAD layer. The frame-level speaker representation is extracted by the deep neural network with differential operations to calculate the dynamic changes between frames, which is more conducive to capturing insignificant changes in the voiceprint. Meanwhile, NeXtVLAD is adopted to split the frame-level features into multiple word spaces before aggregating, and subsequently perform VLAD operations in each subspace, which can significantly reduce the number of parameters and improve performance. Secondly, the margin-based softmax loss function and the few-shot learning-based loss function are proposed to be combined for more discriminative speaker embeddings. Finally, for a fair comparison, the experimental results are performed on Voxceleb-1 showing superior performance of speaker verification system and can obtain new state-of-the-art results.</p> </abstract>

Voice spoofing detection corpus for single and multi-order audio replays

The evolution of modern voice-controlled devices (VCDs) has revolutionized the Internet of Things (IoT) and resulted in the increased realization of smart homes, personalization, and home automation through voice commands. These VCDs can be exploited in IoT driven environments to generate various spoofing attacks, including the chaining of replay attacks (i.e. multi-order replay attacks). Existing datasets like ASVspoof 2017, ASVspoof 2019, and ReMASC contain only first-order replay recordings (i.e. replayed once); therefore, they cannot offer evaluation of anti-spoofing algorithms capable of detecting multi-order replay attacks. Additionally, large-scale datasets like ASVspoof 2017 and ASVspoof 2019 do not capture the characteristics of microphone arrays, which are an essential characteristic of modern VCDs. Therefore, there exists a need for a diverse replay spoofing detection corpus that consists of multi-order replay recordings against bona fide voice samples. This paper presents a novel voice spoofing detection corpus (VSDC) to evaluate the performance of multi-order replay anti-spoofing methods. The proposed VSDC consists of first-order (i.e. replayed once) and second-order replay (i.e. replayed twice) samples against the bona fide audio recordings. We ensured to create a diverse replay spoofing detection corpus in terms of environments, recording and playback devices, speakers, configurations, replay scenarios, etc. More specifically, we used 35 microphones, 25 different recording configurations, and 60 different playback configurations for first- and second-order replays to generate a total of 14,050 samples belonging to 19 speakers. Additionally, the proposed VSDC can also be used to evaluate the performance of speaker verification systems in terms of independent speaker verification. To the best of our knowledge, this is the first publicly available replay spoofing detection corpus comprised of first and second-order replay samples. Experimental results signify the effectiveness of the proposed VSDC in terms of evaluating the performance of anti-spoofing methods under multi-order replay attacks and diverse conditions.

Speaker verification system based on articulatory information from ultrasound recordings

Current state-of-the-art speaker verification (SV) systems are known to be strongly affected by unexpected variability presented during testing, such as environmental noise or changes in vocal effort. In this work, we analyze and evaluate articulatory information of the tongue's movement as a means to improve the performance of speaker verification systems. We use a Spanish database, where besides the speech signals, we also include articulatory information that was acquired with an ultrasound system. Two groups of features are proposed to represent the articulatory information, and the obtained performance is compared to an SV system trained only with acoustic information. Our results show that the proposed features contain highly discriminative information, and they are related to speaker identity; furthermore, these features can be used to complement and improve existing systems by combining such information with cepstral coefficients at the feature level.

Mixture linear prediction Gammatone Cepstral features for robust speaker verification under transmission channel noise

In this paper, we present a Mixture Linear Prediction based approach for robust Gammatone Cepstral Coefficients extraction (MLPGCCs). The proposed method provides performance improvement of Automatic Speaker Verification (ASV) using i-vector and Gaussian Probabilistic Linear Discriminant Analysis GPLDA modeling under transmission channel noise. The performance of the extracted MLPGCCs was evaluated using the NIST 2008 database where a single channel microphone recorded conversational speech. The system is analyzed in the presence of different channel transmission noises such as Additive White Gaussian (AWGN) and Rayleigh fading at various Signals to Noise Ratio (SNR) levels. The evaluation results show that the MLPGCCs features are a promising way for the ASV task. Indeed, the speaker verification performance using the MLPGCCs proposed features is significantly improved compared to the conventional Gammatone Frequency Cepstral Coefficients (GFCCs) and Mel Frequency Cepstral Coefficients (MFCCs) features. For speech signals corrupted with AWGN noise at SNRs ranging from (-5 dB to 15 dB), we obtain a significant reduction of the Equal Error Rate (EER) ranging from 9.41% to 6.65% and 3.72% to 1.50%, compared with conventional MFCCs and GFCCs features respectively. In addition, when the test speech signals are corrupted with Rayleigh fading channel we achieve an EER reduction ranging from 23.63% to 7.8% and from 10.88% to 6.8% compared with conventional MFCCs and GFCCs, respectively. We also found that the combination of GFCCs and MLPGCCs gives the highest performance of speaker verification system. The best performance combination achieved is around EER from 0.43% to 0.59% and 1.92% to 3.88%.

Front-end speech enhancement for commercial speaker verification systems

Commercial speaker verification systems are an important component in security services for various domains, such as law enforcement, government, and finance. These systems are sensitive to noise present in the input signal, which leads to inaccurate verification results and hence security breaches. Traditional speech enhancement (SE) methods have been employed to improve the performance of speaker verification systems. However, to the best of our knowledge, the impact of state-of-the-art speech enhancement techniques has not been analyzed for text-independent automatic speaker verification (ASV) systems using real-world utterances. In this work, our contribution is twofold. First, we propose two deep neural network (DNN) architectures for SE, and we compare the performance of the proposed networks with the existing work. We evaluate the resulting SE networks using the objective measures of perceptual evaluation of speech quality (PESQ) and short-time objective intelligibility (STOI). Second, we analyze the performance of ASV systems when SE methods are used as front-end processing to remove the non-stationary background noise. We compare the resulting equal error rate (EER) using our DNN based SE approaches, as well as existing SE approaches, with real customer data and the freely available RedDots dataset. Our results show that our DNN based SE approaches provide benefits for speaker verification performance.

Prosodic features-based speaker verification using speaker-specific-text for short utterances

Over the past several years, Gaussian mixture model and its variants have been dominant architectures in text-independent and text-dependent speaker recognition field. The recognition accuracy of above-mentioned models declines when experimental utterances' length becomes short in practical application. Presently, Mel-frequency cepstral coefficients are generally used to characterise the properties of the vocal tract and widely applied in speech recognition. In addition, prosodic features, such as pitch and formant, are generally considered to describe the glottal characteristics. However, the efficiency of those approaches remains unsatisfactory. In text-dependent short utterance speaker verification systems, prosodic features can assist to improve the recognition result theoretically. In order to optimise the performance of speaker verification systems under the framework of adapted GMM-UBM, we adopt a variant speaker verification system based on prosodic features, in which a dual judgement mechanism is used in order to integrate vocal tract features with prosodic features. Experimental results showed that the new speech recognition system gives a better consequence.

Frame level sparse representation classification for speaker verification

In this paper, we analyze the application of the sparse representation of frames of the speech signal for the speaker verification. It is lately shown that Sparse Representation Classification (SRC), is promising for speaker recognition. We bring evidence that the frame level sparse representation classification resembles process of speech recognition in human sensory system. Since the recognition of different voices (noises) helps individuals to immediately distinguish between the noise and the original speech signal, a noise aware system was designed. As a principal in the sparse representation, we argued the mutual coherence of the dictionary columns, called dictionary atoms, which is not efficiently considered in the already published SRC base speaker verification researches. To suppress the mutual coherence, we use a dictionary learning method to construct a dictionary with effective atoms. Our proposed Frame Level Sparse Representation Classification (FSRC), provides new insights to the SRC based speaker verification. We demonstrate that, in the SRC based speaker verification, using a dictionary whose atoms are orthogonal can be more extensible than a dictionary whose atoms are highly correlated, and that the mutual coherence suppression is even more effective than imposing strong orthogonality on the dictionary atoms. We consider the performance of state-of-the-art speaker recognition systems and the proposed method on NIST SRE 2004 data. Experimental results show that in comparison to baseline methods, when we have enough amount of information in the registration of targets, the proposed method improves the performance of speaker verification system in noisy conditions.

The Effect of Voice over IP Transmission Degradations on MAP-EM-GMM Speaker Verification Performance

Abstract Despite the growing importance of packet switching systems, there is still a shortage of thorough analyses of VoIP transmission effect on speech and speaker recognition performance. Voice over IP transmission systems use packet switching. There is no guarantee of delivery. The main disadvantage of VoIP is a packet loss which has a major impact on the performance experienced by the users of the network. There are several techniques to mask the effects of a packet loss, referred to as packet loss concealment. In this study, the effect of voice transmission over IP on automatic speaker verification system performance was investigated. The analyzed system was based on MAP-EM-GMM modelling methods. Four various speech codecs of H.323 standard were investigated with special emphasis placed on the packet loss phenomenon and various packet loss concealment techniques.

Feature recovery for noise‐robust speaker verification

Noisy condition is an important extrinsic degradation affecting speaker verification system performance. A feature-recovery approach is proposed to eliminate noise-dependent variability in feature space. A frame of the noisy feature vector is recovered using the information of itself and the neighbour feature vectors. Experiments are conducted on noisy test sets for text-dependent speaker verification tasks and the results indicate that the system can achieve significant performance improvement by using recovered feature vectors.

On robustness of speech based biometric systems against voice conversion attack

Voice conversion (VC) approach, which morphs the voice of a source speaker to be perceived as spoken by a specified target speaker, can be intentionally used to deceive the speaker identification (SID) and speaker verification (SV) systems that use speech biometric. Voice conversion spoofing attacks to imitate a particular speaker pose potential threat to these kinds of systems. In this paper, we first present an experimental study to evaluate the robustness of such systems against voice conversion disguise. We use Gaussian mixture model (GMM) based SID systems, GMM with universal background model (GMM-UBM) based SV systems and GMM supervector with support vector machine (GMM-SVM) based SV systems for this. Voice conversion is conducted by using three different techniques: GMM based VC technique, weighted frequency warping (WFW) based conversion method and its variation, where energy correction is disabled (WFW−). Evaluation is done by using intra-gender and cross-gender voice conversions between fifty male and fifty female speakers taken from TIMIT database. The result is indicated by degradation in the percentage of correct identification (POC) score in SID systems and degradation in equal error rate (EER) in all SV systems. Experimental results show that the GMM-SVM SV systems are more resilient against voice conversion spoofing attacks than GMM-UBM SV systems and all SID and SV systems are most vulnerable towards GMM based conversion than WFW and WFW− based conversion. From the results, it can also be said that, in general terms, all SID and SV systems are slightly more robust to voices converted through cross-gender conversion than intra-gender conversion. This work extended the study to find out the relationship between VC objective score and SV system performance in CMU ARCTIC database, which is a parallel corpus. The results of this experiment show an approach on quantifying objective score of voice conversion that can be related to the ability to spoof an SV system.

Wavelet fuzzy LVQ based speaker verification system

This work evaluates the performance of speaker verification system based on Wavelet based Fuzzy Learning Vector Quantization (WLVQ) algorithm. The parameters of Gaussian mixture model (GMM) are designed using this proposed algorithm. Mel Frequency Cepstral Coefficients (MFCC) are extracted from the speech data and vector quantized through Wavelet based FLVQ algorithm. This algorithm develops a multi resolution codebook by updating both winning and nonwinning prototypes through an unsupervised learning process. This codebook is used as mean vector of GMM. The other two parameters, weight and covariance are determined from the clusters formed by the WLVQ algorithm. The multi resolution property of wavelet transform and ability of FLVQ in regulating the competition between prototypes during learning are combined in this algorithm to develop an efficient codebook for GMM. Because of iterative nature of Expectation Maximization (EM) algorithm, the applicability of alternative training algorithms is worth investigation. In this work, the performance of speaker verification system using GMM trained by LVQ, FLVQ and WLVQ algorithms are evaluated and compared with EM algorithm. FLVQ and WLVQ based training algorithms for modeling speakers using GMM yields better performance than EM based GMM.

Psychoacoustic Model Compensation for Robust Speaker Verification in Environmental Noise

We investigate the problem of speaker verification in noisy conditions in this paper. Our work is motivated by the fact that environmental noise severely degrades the performance of speaker verification systems. We present a model compensation scheme based on the psychoacoustic principles that adapts the model parameters in order to reduce the training and verification mismatch. To deal with scenarios where accurate noise estimation is difficult, a modified multiconditioning scheme is proposed. The new algorithm was tested on two speech databases. The first database is the TIMIT database corrupted with white and pink noise and the noise estimation is fairly easy in this case. The second database is the MIT Mobile Device Speaker Verification Corpus (MITMDSVC) containing realistic noisy speech data which makes the noise estimation difficult. The proposed scheme achieves significant performance gain over the baseline system in both cases.

A Study on Universal Background Model Training in Speaker Verification

State-of-the-art Gaussian mixture model (GMM)-based speaker recognition/verification systems utilize a universal background model (UBM), which typically requires extensive resources, especially if multiple channel and microphone categories are considered. In this study, a systematic analysis of speaker verification system performance is considered for which the UBM data is selected and purposefully altered in different ways, including variation in the amount of data, sub-sampling structure of the feature frames, and variation in the number of speakers. An objective measure is formulated from the UBM covariance matrix which is found to be highly correlated with system performance when the data amount was varied while keeping the UBM data set constant, and increasing the number of UBM speakers while keeping the data amount constant. The advantages of feature sub-sampling for improving UBM training speed is also discussed, and a novel and effective phonetic distance-based frame selection method is developed. The sub-sampling methods presented are shown to retain baseline equal error rate (EER) system performance using only 1% of the original UBM data, resulting in a drastic reduction in UBM training computation time. This, in theory, dispels the myth of “There's no data like more data” for the purpose of UBM construction. With respect to the UBM speakers, the effect of systematically controlling the number of training (UBM) speakers versus overall system performance is analyzed. It is shown experimentally that increasing the inter-speaker variability in the UBM data while maintaining the overall total data size constant gradually improves system performance. Finally, two alternative speaker selection methods based on different speaker diversity measures are presented. Using the proposed schemes, it is shown that by selecting a diverse set of UBM speakers, the baseline system performance can be retained using less than 30% of the original UBM speakers.

Robust Session Variability Compensation for SVM Speaker Verification

This paper presents an enhanced nuisance attribute projection (NAP) method to improve the performance of speaker verification systems in mismatched train and test conditions. Unlike the conventional NAP training method that does not take any scheme to discriminate the source of nuisance, the proposed method quantitatively estimates the source of nuisance based on the statistics of given background speakers' eigenvalues. The estimated values are used for defining a discriminative weight for each of background speakers and selectively including the statistics of between-class scatter or of within-class scatter from them. Through the scheme, we intend to design a more robust projection matrix which involves less speaker-dependent or speaker-intrinsic variability while including more latent nuisance factors beyond the common within-class scatter of backgrounds. Experimental results on the recent NIST SRE evaluations demonstrate that the proposed algorithms produce consistent improvement over the previous NAP approaches.

A segment selection technique for speaker verification

The performance of speaker verification systems degrades considerably when the test segments are utterances of very short duration. This might be either due to variations in score-matching arising from the unobserved speech sounds of short speech utterances or the fact that the shorter the utterance, the greater the effect of individual speech sounds on the average likelihood score. In other words, the effects of individual speech sounds have not been cancelled out by a large number of speech sounds in very short utterances. This paper presents a score-based segment selection technique for discarding portions of speech that result in poor discrimination ability in a speaker verification task. Theory is developed to detect the most significant and reliable speech segments based on the probability that the test segment comes from a fixed set of cohort models. This approach, suitable for any duration of test utterance, reduces the effect of acoustic regions of the speech that are not accurately modelled due to sparse training data, and makes a decision based only on the segments that provide the best-matched scores from the segment selection algorithm. The proposed segment selection technique provides reductions in relative error rate of 22% and 7% in terms of minimum Detection Cost Function (DCF) and Equal Error Rate (EER) compared with a baseline used the segment-based normalization, when evaluated on the short utterances of NIST 2002 Speaker Recognition Evaluation dataset.

열악한 환경에 강인한 화자인증을 위한 위상 기반 특징 추출 기법

화자인증 시스템은 훈련 환경과 인식 환경이 다른 경우 인식 성능이 크게 저하된다. 이러한 훈련과 인식 환경의 불일치는 다양한 잡음과 상이한 채널 환경 때문이다. 본 논문은 화자인증 시스템의 강인성 개선을 위하여 음성신호의 위상에 기반한 특정 추출 기법을 제안한다. 이 방법은 음성신호의 위상으로부터 순시 주파수를 계산하여 대역별로 순시 주파수를 모두 모아 구한 히스토그램으로부터 특징 계수를 추출한다. 이 특징 파라미터를 적용한 결과 조 용한 환경뿐만 아니라 잡음환경 그리고 채널 왜곡 환경에서도 화자인증 시스템의 성능이 개선됨을 알 수 있다. Additive noise and channel distortion strongly degrade the performance of speaker verification systems, as it introduces distortion of the features of speech. This distortion causes a mismatch between the training and recognition conditions such that acoustic models trained with clean speech do not model noisy and channel distorted speech accurately. This paper presents a phase-related feature extraction method in order to improve the robustness of the speaker verification systems. The instantaneous frequency is computed from the phase of speech signals and features from the histogram of the instantaneous frequency are obtained. Experimental results show that the proposed technique offers significant improvements over the standard techniques in both clean and adverse testing environments.

Improving the characterization of the alternative hypothesis via minimum verification error training with applications to speaker verification

Speaker verification is usually formulated as a statistical hypothesis testing problem and solved by a log-likelihood ratio (LLR) test. A speaker verification system's performance is highly dependent on modeling the target speaker's voice (the null hypothesis) and characterizing non-target speakers’ voices (the alternative hypothesis). However, since the alternative hypothesis involves unknown impostors, it is usually difficult to characterize a priori. In this paper, we propose a framework to better characterize the alternative hypothesis with the goal of optimally distinguishing the target speaker from impostors. The proposed framework is built on a weighted arithmetic combination (WAC) or a weighted geometric combination (WGC) of useful information extracted from a set of pre-trained background models. The parameters associated with WAC or WGC are then optimized using two discriminative training methods, namely, the minimum verification error (MVE) training method and the proposed evolutionary MVE (EMVE) training method, such that both the false acceptance probability and the false rejection probability are minimized. Our experiment results show that the proposed framework outperforms conventional LLR-based approaches.

Mandarin-Sichuan dialect bilingual text-independent speaker verification using GMM

Due to the mismatch between mandarin and Sichuan dialect in training and test stages,the performance of speaker verification system degrades dramatically.To solve this problem,a combined Gaussian Mixture Model(GMM),which is trained by proportional pooling mandarin and Sichuan dialect,was presented in this paper.Compared with the Gaussian mixture model trained solely using mandarin/Sichuan dialect,the combined Gaussian mixture model described the characteristic of speaker from both mandarin and Sichuan dialect.Experiments on a self-built mandarin-Sichuan dialect speech database demonstrate that the introduced combined Gaussian mixture model is more robust for speech mismatching between mandarin and Sichuan dialect.A proper proportion between pooling mandarin and Sichuan dialect speech was also provided.

Feature mapping and recuperation by using elliptical basis function networks for robust speaker verification

The performance of speaker verification systems is often compromised under real-world environments. For example, variations in handset characteristics could cause severe performance degradation. This paper presents a novel method to overcome this problem by using a non-linear handset mapper. Under this method, a mapper is constructed by training an elliptical basis function network using distorted speech features as inputs and the corresponding clean features as the desired outputs. During feature recuperation, clean features are recovered by feeding the distorted features to the feature mapper. The recovered features are then presented to a speaker model as if they were derived from clean speech. Experimental evaluations based on 258 speakers of the TIMIT and NTIMIT corpuses suggest that the feature mappers improve the verification performance remarkably.